This application addresses PQB-5: How does the order in which mutations or epigenetic changes occur alter cancer phenotypes or affect the efficacy of targeted therapies? Recently published genome sequencing data for acute myeloid leukemia (AML) from The Cancer Genome Atlas (TCGA) has demonstrated that >80% of AML clones contain at least two mutations, and 44% include mutations in DNA-methylation-related genes, supporting that epigenetic regulation plays a key role in AML development. It currently remains completely unknown how the order of acquisition of these mutations or epigenetic changes impacts AML phenotype and response to targeted therapy. Studying combinatorial acquisition of mutations directly in humans is challenging due to the necessity of sequential sampling of tumors and prospective identification of pre-tumorigenic cell populations. To meet this challenge, we have devised an in vitro-based modeling system to accurately depict the consequences of sequential acquisition of mutations in AML. Using this system, we will sequentially and combinatorially induce common, recurrent mutations in the DNA methylation machinery (DNMT3A and TET2) and a frequently associated activating mutation in FLT3 (FLT3-ITD). We will determine how the order of acquisition of mutations in DNMT3A, TET2 and FLT3 alters leukemia phenotype, and establish signatures of enhancer activity that are predictive of the identity of the initiating mutation. We will test our hypothesis that active enhancer regions can serve as key biomarkers for the identification of initiating mutations in AML and that these predict response to the targeted DNA methylation inhibitor decitabine (DAC). Our results will have broad impact on cancer research by direct demonstration that the order of acquisition of mutations (1) alters cancer phenotype, (2) can be predicted by epigenetic biomarkers, and (3) is a key factor for therapeutic decisions.